TLDR - my first efforts at creating AI videos can be found here (I have no plans to post videos on Youtube): https://umd.net/profile/i/wetmaxiskirts/section/videos

Until now, I hadn't delved into the realm of AI videos as I thought I would never had the time or energy needed to produce them. As it turned out, it was actually so much easier than I ever thought it might be.

Inspired by a few others here, I decided to give Whisk/Veo a go and found that the process of turning an image created using a standard text prompt into a video is actually very simple. You just click "Animate" and add a further prompt to provide some guidelines on the animation you'd like to see and also what you want to be said in the video.

Well the theory is simple but, as with all AI engines, the actuality is a bit different! Firstly, videos can only be created for images with a 16:9 landscape aspect ratio, so you have to come up with image content that fits comfortably into that.

Secondly, as most people here will know, the videos are limited to 8 seconds in length, so the required action and the speech have to fit into that timeframe.

Thirdly, the most frustrating of all - most of the attempts at animating an image fail with "error creating video". As it takes up to 2 minutes to generate a video, this message is very frustrating when it occurs right at the end of the process. I was quite lucky at first, but it soon became the case that the vast majority of attempts I made crashed and burned in this way. My animation prompts are pretty inocuous, as befits my preferred type of wetlook, but the failures just keep coming.

Rather than assuming that my animation prompt is always the culprit, I'm wondering if it's the case that some images simply can't be turned into videos no matter what I put in the prompt? As the number of videos is limited to 50 per month, I don't want to waste my allocation by experimenting too much. If anyone has any top tips for avoiding "error creating video", I'd be very grateful for them!

Anyway, onto the good bit. Quite simply being able to bring an image to "life", even if it's just for 8 seconds, gives me a fantastic buzz. Seeing the "play video" symbol, rather than "error creating video", leads to an immense sense of excitement and anticipation, knowing that something approaching what's specified in my animation prompt is about to unfold before my eyes! Obviously there are a few glitches to be spotted in some of the videos because that's how AI still is as we all know. However, the job that Veo makes of piecing everything together is quite remarkable, to the extent That I was absolutely gobsmacked when watching some of these videos for the first time.

It's still very early days for me and I need to learn more about how best to define both the actions and the speech, but it's also nice to see what the AI comes up with itself. The second of the "Boardgames in the Rain" videos was generated with no speech guidance from me and yet the AI nevertheless generated a suitable comment from a wetlook perspective!

Speaking of videos there are 6 of them, each featuring 2 different scenes on the same theme, so there are 12 individual 8 second videos in all. I've done it this way because videos on UMD have to be at least 15 seconds long. The 12 images below are what the videos have been created from and I was completely enthused and inspired by being able to give them "life"!

They're all great, especially the dialogue.